A repetitive mutation and selection system for bacterial evolution to increase the specific affinity to pancreatic cancer cells.

Masaki Osawa,Alvaro Galli

doi:10.1371/journal.pone.0198157

Abstract

It is difficult to target and kill cancer cells. One possible approach is to mutate bacteria to enhance their binding to cancer cells. In the present study, Gram-negative Escherichia coli and Gram-positive Bacillus subtilis were randomly mutated, and then were positively and negatively selected for binding cancer vs normal cells. With repetitive mutation and selection both bacteria successfully evolved to increase affinity to the pancreatic cancer cell line (Mia PaCa-2) but not normal cells (HPDE: immortalized human pancreatic ductal epithelial cells). The mutant E. coli and B. subtilis strains bound to Mia PaCa-2 cells about 10 and 25 times more than to HPDE cells. The selected E. coli strain had mutations in biofilm-related genes and the regulatory region for a type I pilus gene. Consistent with type I pili involvement, mannose could inhibit the binding to cells. The results suggest that weak but specific binding is involved in the initial step of adhesion. To test their ability to kill Mia PaCa-2 cells, hemolysin was expressed in the mutant strain. The hemolysin released from the mutant strain was active and could kill Mia PaCa-2 cells. In the case of B. subtilis, the initial binding to the cells was a weak interaction of the leading pole of the motile bacteria. The frequency of this interaction to Mia PaCa-2 cells dramatically increased in the evolved mutant strain. This mutant strain could also specifically invade beneath Mia PaCa-2 cells and settle there. This type of mutation/selection strategy may be applicable to other combinations of cancer cells and bacterial species.

Highlights

A potential approach to cancer treatment is to target the cancer cells with toxic reagents
Bacteria randomly mutated with UV irradiation were subjected to negative selection with immortalized human pancreatic ductal epithelial (HPDE) cells and positive selection with a pancreas ductal adenocarcinoma cell lines (MIA PaCa2 cells) (Fig 1 and see materials and methods)
To concentrate bacteria that bind to the cancer cells, three rounds of negative and positive selection without UV irradiation were applied after each UV irradiation (Fig 1)

Summary

Introduction

A potential approach to cancer treatment is to target the cancer cells with toxic reagents. One possibility to target cancer cells would be a probe that recognized the quantitative difference in surface epitopes, perhaps by interacting weakly with multiple types of molecules that are expressed more. One interesting observation is that various B. subtilis strains have largely different affinities for mucin, matrigel and a heterogeneous human epithelial colorectal adenocarcinoma cell line (Caco-2 cells) [6] This suggests that random mutations might affect the bacterial surface and alter their binding to cell surface antigens. One interesting approach has been selection of random aptamers reported by cyclic negative and positive selection for binding to secreted materials from non-cancer pancreatic epithelial cells vs pancreatic ductal adenocarcinoma cells [18]. The repetitive mutation/selection system produced bacteria that bind to the pancreatic ductal adenocarcinoma cells

Methods

Results

G AG TAATTTCTTTCGTCGTAGTAATTGCATGAGTT

Discussion